Why 8 Bits? The Short, True History
Nothing in physics demands 8. Early machines grouped bits in sixes, sevens and nines — a "byte" originally meant whatever chunk a machine chewed per step. The consolidation came with IBM's System/360 in 1964, which committed to 8-bit bytes: enough to hold any character of the era's 7-bit ASCII with a bit to spare, and a power of two, which binary addressing handles elegantly. The 360's dominance made 8 the industry default, microprocessors inherited it, and by the 1970s the 8-bit byte was simply what "byte" meant. The standards later made it official — IEC 80000-13 defines the byte as 8 bits.
One fossil of the ambiguous era survives: networking RFCs say "octet" instead of "byte" precisely because, when the internet's protocols were written, a "byte" wasn't yet reliably 8 bits everywhere. Same unit, lawyer-proof name.